Hybrid-Bonded Image Sensor Architecture for Maximizing Photodiode Area

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

CMOS image sensors face challenges in maximizing photodiode area due to the need for transistor and circuitry space, especially in compact devices, which affects sensitivity as pixel count increases.

Innovation Solution

A hybrid-bonded image sensor architecture where a photodiode die with macrocells is bonded to a supporting circuitry die with supercells, allowing for optimized photodiode area usage and separate optimization of photodiode and CMOS circuitry materials and processing, with amplification and digitization occurring through common source and differential amplifiers.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of moving object

If photodiode and circuitry are integrated on the same die, then device complexity is reduced and manufacturing is simplified, but photodiode area is reduced which decreases sensitivity

Engineering Contradiction:
Improvephotodiode areaVSAvoiddevice complexity
Core Design Contradiction:
Area of moving objectVSDevice complexity

Solution Approach 1:

The image sensor is divided into two separate dies: a photodiode die containing only photodiodes and a circuitry die containing all transistors and supporting circuitry. This segmentation allows the photodiode die to be optimized for light sensing with maximum area, while the circuitry die handles all electronic functions, thereby resolving the area-complexity contradiction.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Bond contacts serve as intermediaries connecting the photodiode die to the circuitry die. These bond contacts enable electrical connection between the photodiodes and the supporting circuitry while allowing the two functional blocks to be physically separated, thus maximizing photodiode area without losing electrical connectivity.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If pixel count is increased in compact devices, then image quality and resolution are improved, but photodiode area per pixel is reduced which affects sensitivity

Engineering Contradiction:
Improveimage sensor performanceVSAvoidphotodiode area
Core Design Contradiction:
Measurement precisionVSArea of moving object

Solution Approach 1:

By segmenting the sensor into dedicated photodiode and circuitry dies, each pixel's photodiode can be maximized in area on the photodiode die, while the circuitry is efficiently packed on the separate circuitry die. This allows higher pixel counts in compact form factors without sacrificing individual photodiode area and sensitivity.

Inventive Principle:
Principle #1Segmentation

3Reliability

If photodiode area is maximized, then sensitivity is enhanced, but space for transistors and circuitry is reduced

Engineering Contradiction:
ImprovesensitivityVSAvoidcircuitry space
Core Design Contradiction:
ReliabilityVSArea of stationary object

Solution Approach 1:

The sensor is segmented into a photodiode die optimized for sensitivity with maximum photodiode area, and a separate circuitry die optimized for housing transistors and supporting circuitry. This eliminates the area trade-off by providing dedicated space for each function on separate dies.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The solution moves from a two-dimensional planar integration challenge to a three-dimensional stacked architecture using hybrid bonding. By utilizing the vertical dimension through die stacking, the photodiode die can be fully dedicated to light sensing while the circuitry die handles electronic functions, effectively adding a spatial dimension to resolve the area conflict.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach maximizes photodiode area, enhancing sensitivity and enabling larger pixel counts in compact devices by separating photodiode and circuitry fabrication, thus improving image sensor performance.

Implementation Method 1

a photodiode die with multiple macrocells, where each macrocell has at least one photodiode

Methodology Applied
Scientific EffectPhotoelectric effect: Photoelectric Effect

Implementation Method 2

at least one common source amplifier adapted to receive signal from the bond contact of a selected macrocell unit of the supercell

Methodology Applied
Scientific EffectTransistor amplification:

Data Source

PatentUS10375338B2Two stage amplifier readout circuit in pixel level hybrid bond image sensors
Publication Date: 2019.08.06 OMNIVISION TECHNOLOGIES INC
  • US10375338B2 patent drawing
  • US10375338B2 patent drawing
  • US10375338B2 patent drawing

AI summary

A hybrid bonded image sensor has a photodiode die with macrocells having at least one photodiode and a bond contact; a supporting circuitry die with multiple supercells, each supercell having at least one macrocell unit having a bond contact coupled to the bond contact of a macrocell of the photodiode die. Each macrocell unit lies within a supercell and has a reset transistor adapted to reset photodiodes of the macrocell of the photodiode die. Each supercell has at least one common source amplifier adapted to receive signal from the bond contact of a selected macrocell unit of the supercell, the common source amplifier coupled to drive a column line through a selectable source follower. In embodiments, the common source amplifiers of several supercells drive the selectable source follower through a distributed differential amplifier.